Séance Séminaire

Séminaire de Probabilités et Statistique

Functional linear models with partially observed data

Dealing with missing values is an important issue in data observation or data recording process. In this work, we consider a functional linear regression model with partially observed covariate and missing values in the response at first time. We use a reconstruction operator that aims at recovering the missing parts of the explanatory curves, then we are interested in regression imputation methods of missing data on the response variable, using functional principal component regression to estimate the functional coefficient of the model. In the second part, we are interested in a function-on-function linear model in which the response and the covariate are partially observed curves. We consider two strategies for dealing with the missing part of the response. The first one consists in a reconstruction in the same way as for the covariate. The second one uses regression imputation. Once the dataset is reconstructed, we study the asymptotic behavior of the prediction error. The methods are compared from a theoretical and a practical point of view.

Séminaire en salle 109, également retransmis sur zoom : https://umontpellier-fr.zoom.us/j/94087408185